106 BOTANICAL GAZETTE ; [AUGUST 
me to be very unsafe to infer that any particular one is so, without 
independent proof thereof. 
6. Monstrosities or abnormalities —So far as concerns inherited 
abnormalities, it must be true that in the long run the progressive 
changes must have much exceeded the reversions, else evolution 
from the simpler ancestors to the more complex descendants of 
today could not have occurred. This argument cannot be applied 
in the same way to non-heritable abnormalities, though there is 
no obvious reason why they should follow any different law of 
probability. By themselves, abnormalities afford evidence of 
little weight, since it is impossible to say whether they represent 
reversions or other chance variations. 
Conclusions 
We may now attempt to apply these criteria of evidence to the 
arguments that have been offered by the various supporters of each 
of the theories. So far as concerns the lycopod theory and the 
cordaitean theory, it is readily seen that each of them is founded 
on certain more or less striking resemblances. The evidence, 
then, is valid so far as the principles are concerned. The weakness 
of each theory lies in the necessity of certain more or less plausible 
explanations that must be accepted before the resemblances are 
evident. The lycopod theory must explain away the very appar- 
ent difference in the stelar structures of the two groups and must 
show how the pine cone has been evolved from the simple cone of a 
club moss. Without repeating what has already been set forth in 
other parts of this paper, let it suffice to recall the very large number 
of points of difference that must be explained away and the com- 
paratively few points of likeness relied on to establish a relationship. 
On the contrary, the points of resemblance between araucarians 
and Cordaitales are numerous and striking. The points of differ- 
ence are few, and for the most part more easily explained away than 
those that confront the preceding theory. The geological record 
appears also to favor the cordaitean theory, for none of the fossil 
forms are known to approach lycopods in any character more closely 
than do the modern forms. In fact, it would appear from our pres- 
ent knowledge that the fossil forms were less like them than the 
